Repository navigation
Releases: stackhpc/ansible-slurm-appliance
Release list
v2.30.0
What's Changed
This update is provided primarily to address CVE-2026-99999 (PwnEscape). See PR1075 below.
Security and updates
- Fix GHSA-6q48-6c7w-9gw2 by @wtripp180901 in #1069
- Update OS packages for CVE-2026-99999 by @sjpb in #1075
General enhancements and fixes
- Support Slurm REST API (slurmrestd) by @elelaysh in #1056
- Add sbatch options for hpctests by @sjpb in #1052
- Fix missing volumes info in hosts.yml for additional_nodes by @sjpb in #1071
- Add psi=1 to kernel command-line by @elelaysh in #931
- Bump stackhpc/prometheus-community-ansible to 0.30.1+stackhpc.1 by @elelaysh in #1067
StackHPC CI and development tools
Full Changelog: v2.29.0...v2.30.0
Images
Two new images are available:
- RL8: openhpc-RL8-261026-1234-b24dffa1
- RL9: openhpc-RL9-261026-1234-b24dffa1
v2.29.0
What's Changed
This update is provided primarily to address a CVE with the kernel vhost module - see CVE-2026-74580 / PR1063 below.
Security and updates
- Add systemtap role by @wtripp180901 in #1035
- Block vhost module for CVE-2026-74580 by @sjpb in #1063
- Sentence CVE-2026-78663 / GO-2026-6612 by @sjpb in #1064
General enhancements and fixes
- Fix Open OnDemand DEX mode broken since ood.osc 4.1.6 by @elelaysh in #1057
- Bump CUDA to 13.4.3 by @priteau in #1062
StackHPC CI and development tools
Full Changelog: v2.28.0...v2.29.0
Images
Two new images are available:
- Rocky Linux 8: openhpc-RL8-261015-1924-67853625
- Rocky Linux 9: openhpc-RL9-261015-1924-67853625
v2.28.0
What's Changed
This release is primarily provided to address kernel network module CVEs - see PR 1055 below.
Security and updates
General enhancements and fixes
- Skip prometheus install in configure mode by @sjpb in #1051
- Run gpuburn as part of hpctests by @elelaysh in #975
StackHPC CI and development tools
- Bump grype-analyse for package info and unused ignore rule fixes by @sjpb in #1053
- Scan images in main branch every night and send Slack message on failure by @elelaysh in #1039
- Deal with false-positives in scan for CVE 2026 90711 by @sjpb in #1059
Full Changelog: v2.27.0...v2.28.0
Images
Two new images are available:
- Rocky Linux 8: openhpc-RL8-261003-1009-c76116df
- Rocky Linux 9: openhpc-RL9-261003-1009-c76116df
v2.27.0
What's Changed
This release is provided primarily to support aes256k cephx keys to allow mitigation of CVE-2025-30156 for clusters with CephFS/Manila mounts.
Security and updates
- Update RL9 Ceph packages to Squid to support aes256k cephx keys by @MoteHue in #1048
- hpctests minor fixes by @elelaysh in #1041
General enhancements and fixes
StackHPC CI and development tools
Full Changelog: v2.26.0...v2.27.0
Images
Two new images are available:
- Rocky Linux 8: openhpc-RL8-260930-1437-38a6351e
- Rocky Linux 9: openhpc-RL9-260930-1434-38a6351e
Note syft sboms are available at the same link with matching file names.
v2.26.0
What's Changed
Security and updates
- Block ah6, tun, pppoe, sctp, sctp_diag kernel modules due to CVEs by @elelaysh in #1045
- Bump CUDA to 13.4.2 by @priteau in #1043
General enhancements and fixes
- Make slurm-controlled reboots drain nodes first and run in batches by @sjpb in #1032
- Small fixes to weka role by @elelaysh in #1044
- Bump openhpc role to validate SlurmdSpoolDir changes by @sjpb in #1042
Full Changelog: v2.25.0...v2.26.0
Images
Two new images are available:
- Rocky Linux 8: openhpc-RL8-260922-1455-98b677bc
- Rocky Linux 9: openhpc-RL9-260922-1455-98b677bc
Note syft sboms are available at the same link with matching file names.
v2.25.0
What's Changed
Warning
(Added 16/09/2026)
In v2.20 the default slurmd spool directory (SlurmdSpoolDir) was changed from /var/spool/slurm to the Slurm default of /var/spool/slurmd. Applying this change via the site.yml playbook terminates running jobs. If you have not yet upgraded to v2.20 or later, and plan to upgrade with jobs running (e.g. using slurm-controlled rebuild), you should explicitly set the previous default via an override:
# environments/site/inventory/group_vars/all/openhpc.yml:
openhpc_config_extra:
SlurmdSpoolDir: /var/spool/slurmIf you will not have jobs running during the upgrade you should not apply this override.
This release updates slurm to 25.05.9 for Rocky Linux 8 and 25.11.8 for Rocky Linux 9 to address various CVEs. This is a minor update for Slurm on both OSes if upgrading from appliance version 2.20 or later.
It also adds experimental functionality to limit loaded kernel modules for security - see docs/experimental/modulejail.md.
Security and updates
General enhancements and fixes
- Fix proxy defaults for ansible-core 2.16.14 by @technowhizz in #1025
- Support baremetal ConnectX-7 infiniband in nhc role by @elelaysh in #988
- Migrate Grafana dashboards, upgrade Grafana to v12.4.4, grafana-opensearch-datasource to 2.33.1 and Slurm exporter to 0.23.0 by @wtripp180901 in #987
- Bump openhpc role for DB timeout and validation fixes by @sjpb in #1030
- Fix nhc export task undefined variable in templating by copying live config by @elelaysh in #1029
- Fix opensearch role idempotence regarding owner/group by @elelaysh in #1028
- Remove test CA cert from release images by @sjpb in #1027
- Bump CUDA to 13.4.1 and NVIDIA driver to 615.71.09 by @priteau in #1033
- Update Slurm to 25.05.9 (RL8) and 25.11.8 (RL9) by @elelaysh in #1036
StackHPC CI and development tools
- Change README workflow badge to use correct workflow by @sjpb in #1026
- Run syft in the packer VM and reference sbom in openstack image property by @elelaysh in #1020
- Fix sbom.zip location in Upload Image and SBOM to S3 by @elelaysh in #1037
- Fix remaining space in some values in s3-image-sync by @elelaysh in #1038
Full Changelog: v2.24.0...v2.25.0
Images
Two new images are available:
- Rocky Linux 8: openhpc-RL8-260914-0843-1e1e0995
- Rocky Linux 9: openhpc-RL9-260914-0843-1e1e0995
Note syft sboms are provided with matching file names.
v2.24.0
What's Changed
This release includes a fix for CVE-2026-53264 on RockyLinux 9 (RockyLinux 8 is not vulnerable).
Security and Updates
- Add tool for live patching CVEs by @wtripp180901 in #1005
- Bump repo timestamps to fix CVE-2026-53264: net/sched race by @elelaysh in #1023
General enhancements and fixes
- Switch to include_role for roles in prometheus collection by @jovial in #1011
- Fix ansible-init scontrol update check by @MoteHue in #1016
- Bump CUDA to 13.3.1 and NVIDIA driver to 610.57.04 by @priteau in #1015
- Add patch to stackhpc images by @elelaysh in #1013
- Include git commit in /var/lib/image/image.json by @elelaysh in #1014
- Avoid rebooting nodes from site.yml by @sjpb in #1018
- Revert "Mitigate ssh-keysign-pwn exploits (#965)" by @jovial in #1003
- Fix logic when with-mkhomedir is not currently configured by @jovial in #1012
- Update to lustre 2.15.8 & test build in RL8 CI by @sjpb in #984
- Fix pulp sync python version for RL8 by @sjpb in #1022
- Support WEKA filesystem client mount by @elelaysh in #996
StackHPC CI & development tools
- dev/delete-cluster.py: purge volume snapshots by @elelaysh in #1019
- Disable nightly-cleanup in ARCUS cloud by @elelaysh in #1021
Full Changelog: v2.23.0...v2.24.0
Images
Two new images are available:
- RockyLinux 8: openhpc-RL8-260818-0723-3f645271
- RockyLinux 9: openhpc-RL9-260818-0723-3f645271
v2.23.0
v2.22.0
What's Changed
- Support running HPC tests in a Slurm reservation by @priteau in #1002
- Bump pulp snapshots to fix CVE-2026-64600 by @jovial in #1006
Images
Two new images are available:
- RockyLinux 8: openhpc-RL8-260724-1046-3267c585
- RockyLinux 9: openhpc-RL9-260724-1045-3267c585
Full Changelog: v2.21.2...v2.22.0
v2.21.2
What's Changed
Security and Updates
Full Changelog: v2.21.1...v2.21.2
Images
A new image is available:
- RockyLinux 8: openhpc-RL8-260717-1018-bcd59178
See v2.21 for the unchanged RockyLinux 9 image.